Summary
Alphabet Inc. shares declined after reports that Google is months behind schedule on releasing Gemini 3.5 Pro, a flagship large language model. The setback is attributed to additional time taken to improve the model’s capabilities, with a particular emphasis on coding. The delay has prompted frustration inside the company and has drawn attention to competitive dynamics as other firms roll out models that currently surpass Google’s offerings in code-related tasks.
Details
Shares of Alphabet fell 2.3% on Thursday afternoon following news that the company is delayed in delivering Gemini 3.5 Pro. According to people familiar with the situation, the extra time being taken is focused on strengthening the model’s abilities in areas such as programming and code generation. That work follows an earlier update late last month in which Google adjusted the training data used for Gemini to try to improve coding performance, but those updates reportedly produced disappointing results.
Inside Google, engineers, AI researchers and managers have expressed frustration at the pace of progress. Many of those cited voiced concern that Google could cede ground to competitors as firms such as Anthropic and OpenAI (and other industry players) produce models that are judged to exceed Gemini’s current capabilities in coding tasks.
A Google spokesperson said the company is shipping rapidly across a wide range of models while working to keep them cost-effective for customers. The spokesperson added that Google is testing 3.5 Pro, an upgraded Flash model, and other models with partners. The company is also engaged with the U.S. government on model testing and broader safety frameworks.
The report highlighting the delay noted that Google’s internal structure - with multiple layers of stakeholders involved in preparing models for release and in integrating AI across a broad portfolio of products including search, maps and YouTube - can contribute to slower rollouts.
Key points
- Alphabet shares fell 2.3% after a report that Google is months behind schedule on Gemini 3.5 Pro.
- The delay is driven by extra work to improve the model’s coding capabilities; prior training-data updates produced disappointing results.
- Google says it is testing upgraded models with partners and engaging with the U.S. government on testing and safety frameworks; the company emphasizes rapid shipping and cost-effectiveness.
Risks and uncertainties
- Competitive risk: The company faces pressure if rivals’ models surpass Gemini in code-writing performance, which could affect product competitiveness across Google’s services.
- Execution risk: Multiple internal stakeholders and the complexity of integrating models across search, maps and YouTube may slow development and releases.
- Performance risk: Recent updates to the model’s training data intended to improve coding ability produced disappointing results, leaving uncertain timelines for achieving target capability.
Context limitations
If further detail is required on testing timelines, internal decision processes or specific performance metrics, the available information is limited to the company statements and the reporting cited above.